GSoC participation roadmap of Echedey

[echedey-ls]

Hi everybody!

After a welcoming meeting I had with this year GSoC mentors, @kandersolar and @AdamRJensen, and the other two participants, @RDaxini and @IoannisSifnaios, we were recommended to have a tracking issue of the milestones for each one of us. Here's mine.

Objectives

The proposed changes in this roadmap expect to improve the availability of models so users can include and improve their PV systems simulations. Topics are:

• Shade calculation and impact
• Spectra analysis utilities
• Bifacial modelling of non-uniform irradiance
• Agrivoltaics useful radiation calculations

Roadmap

• Martinez shading losses model

  • Depends on merging Shaded Fraction on Sloped Terrains - PR1725 partial continuation #1962
  • Shading model that takes into account the bypass diodes and layout of modules #2063
  • Martinez shading factor #2070

• Spectra analysis

  • Extend standard spectra database
    • Add full ASTM G-173-03 tables #1963
    • Add extraterrestrial and direct spectra of ASTM G173-03 standard spectrum (AM1.5) #2039
  • Add Spectral response <> Quantum Efficiency conversion functions
    • Spectral responsivity and quantum efficiency conversion functions #2040
    • Quantum efficiency & spectral response conversion funcs #2041
  • Add spectral responses of materials
    • Spectral response of other materials #2037
    • [DRAFT] Add more spectral responses #2038

• Non-uniform irradiance mismatch losses model

  • Reduced order electrical mismatch model #1541
  • NREL non-uniform irradiance mismatch loss for bifacial modules #2046
  • NREL mismatch loss addendum #2147

• Photosynthetically Active Radiation models

  • Decomposition: fraction of diffuse (Spitters relationship)
    • Photosynthetically active radiation decomposition model #2047
    • Agrivoltaics - PAR diffuse fraction model #2048
  • Diffuse and direct shading factors - 3D geometry calculation of shades via geometry and integration.
    I'm afraid this one is doing PVsyst-like calculations all over the 3D and 2D space, so it's gonna take some time to sort out all the difficulties and integrate it here in pvlib.
    Here is the paper meanwhile.
    • Direct and diffuse shading factors - spatial approach (for Photosynthetically Active Radiation) #2069
    • [Feature Preview] Shadow casting on and by flat surfaces #2106

I want to open issues and PRs as I work on them. If I can manage so many things at the same time is something I have to figure out, not you - so don't hesitate to give as much feedback as you want! 😆

I'm looking forward to use my knowledge to improve this package, but as the inexperienced student I am, I can't do it without y'all, so I want to hear from you in either issues or PRs 🤙

Other PRs that I have worked on during this GSoC period

• Fix silently ignoring file path in pvsystem.retrieve_sam when name is provided #2020
  Opened on May 1st, may be considered part of the bonding period.
• More docs build fixes #2030
• Remove read_srml_month_from_solardat #2061 (closed as duplicate)
• Fix port of old-precision-decimal to present-atol in tests #2082
• No more default nones #2084
• Bump docs infraestructure #2112
• Remove warnings in docs #2128
• [MAINT]: Remove deprecated pvlib.atmosphere.first_solar_spectral_correction #2131
• On other repos
  • Solarfactors notebook openpvtools/pv-foss-engagement#9
  • Fix mismatch fit3, add mismatch fit2 NREL/bifacial_radiance#520

Cross-referencing:

• My GSoC mail thread
• Personal blog
• GSoC project page

[cwhanse]

Thinking about the API for the agrivoltaics functions, modeling an agrivoltaic system involves a modeling chain that handles both the photovoltaic and crop production. Some steps in that chain are in common, e.g., separating GHI into diffuse and direct, but more likely, many steps are in parallel, e.g. calculate cell temperature and calculate air temperature for the crops. The same function won't be used for both steps.

Two options suggest themselves:

1. put agrivoltaics functions into the code module for the corresponding photovoltaic modeling step, i.e., irradiance, temperature, unless there's no corresponding photovoltaic step, in which case the function goes into an agrivoltaic module, e.g., interface function to crop production model libraries.
2. put all agrivoltaics functions into a separate module(s) the scope of which is only agrivoltaics.

I lean to option 1, since I prefer to look at agrivoltaics as a variation of the core pvlib workflow. I have the same view of floating PV modeling.

If we agree on option 1, then the PAR diffuse fraction function would go with the rest of the GHI separation functions, into pvlib.irradiance.

[echedey-ls]

Right now the opinions are the same for pvlib.spectrum and pvlib.irradiance, so anyone visiting this conversation is encouraged to leave their thoughts at #2047! 😃

@cwhanse I agree with your suggestion. I'd like you to have a look, if possible, to the PAR shading factors paper since they are more oriented to a geometrical 3D space approach, and as of now, I don't think pvlib has any functionality regarding that. In any case, I will open an issue in short so we can move that conversation there.

[echedey-ls]

This GSoC has come to its end, so this issue can be safely closed :D

❣️ Thank you very much to everybody, specially the mentors @kandersolar & @AdamRJensen and fellow contributors @RDaxini & @IoannisSifnaios for the funny meetings, and all the maintainers and community members involved in my PRs and discussions (@cwhanse , @adriesse , @mikofski ...)

I will try to continue contributing to pvlib in the future as well 🫡